import dagster as dg from dagster_snowflake import SnowflakeResource from src.utils.snowflake.constants import Warehouse from src.utils.snowflake.query import load_query from src.utils.snowflake.partition_utils import parse_hourly_partition_key @dg.asset_check( asset="rds_generated_lyrics", description="Verify ID uniqueness in partition", blocking=False ) def check_generated_lyrics_id_uniqueness( context: dg.AssetCheckExecutionContext, snowflake: SnowflakeResource ) -> dg.AssetCheckResult: """Check that all IDs in the partition are unique.""" partition_key = context.op_execution_context.partition_key p_date, p_hour = parse_hourly_partition_key(partition_key) with snowflake.get_connection() as conn: cursor = conn.cursor() cursor.execute(load_query( "src/utils/snowflake/queries/use_warehouse.sql", params={"warehouse": Warehouse.SUNO_PROD_RDS_HOURLY_MEDUIM.value} )) cursor.execute(f""" SELECT COUNT(*), COUNT(DISTINCT ID), COUNT(*) - COUNT(DISTINCT ID) FROM SUNO_PROD.PROD.RDS_GENERATED_LYRICS WHERE p_date = '{p_date}' AND p_hour = {p_hour} """) total, distinct, dupes = cursor.fetchone() or (0, 0, 0) passed = dupes == 0 return dg.AssetCheckResult( passed=passed, description=f"{'All IDs unique' if passed else f'{dupes} duplicate IDs found'}", metadata={ "total_rows": total, "distinct_ids": distinct, "duplicates": dupes } ) @dg.asset_check( asset="rds_generated_lyrics", description="Verify row count > 0", blocking=False ) def check_generated_lyrics_row_count( context: dg.AssetCheckExecutionContext, snowflake: SnowflakeResource ) -> dg.AssetCheckResult: """Check that the partition has data (row count > 0).""" partition_key = context.op_execution_context.partition_key p_date, p_hour = parse_hourly_partition_key(partition_key) with snowflake.get_connection() as conn: cursor = conn.cursor() cursor.execute(load_query( "src/utils/snowflake/queries/use_warehouse.sql", params={"warehouse": Warehouse.SUNO_PROD_RDS_HOURLY_MEDUIM.value} )) cursor.execute(f""" SELECT COUNT(*) FROM SUNO_PROD.PROD.RDS_GENERATED_LYRICS WHERE p_date = '{p_date}' AND p_hour = {p_hour} """) row_count = cursor.fetchone()[0] passed = row_count > 0 return dg.AssetCheckResult( passed=passed, description=f"Partition has {row_count} rows (expected > 0)", metadata={"row_count": row_count} ) asset_checks = [ check_generated_lyrics_id_uniqueness, check_generated_lyrics_row_count ] __all__ = ["asset_checks"]